This is for anyone who currently host their WordPress or ClassicPress in a virtual private server such as DigitalOcean, Linode, or any other VPS providers. Any Linux user with knowledge of command line can perform backups and restoration tasks. You must be familiar with Linux and you know how to connect to your production server via SSH.
Plus, this article assumes that you have installed and configured WordPress in your VPS server. In addition, this article assumes you can perform basic database administration tasks such as adding a new database along with creating a new user for WordPress or ClassicPress. If your hosting provider provides managed WordPress or ClassicPress hosting, then this article may not apply to you. And because this article is for those who currently run a VPS server, I am going to have to assume that you have some hands-on experience with the Linux command line. This article need not apply to non-technical Linux, Mac, and Windows users. When I say non-technical Linux users, I'm talking about those who wanted to get away from Windows or Mac and simply wanted to use Linux just to browse the Internet and not deal with the command line.
Last, but not least, I am also going to assume that you know how to configure your Apache server as well. Both Apache and NGINX (pronounced Engine-X) configurations won't be covered here, including backing up and copying certificates that you get from your hosting provider.
You have your own VPS server that is running ClassicPress. Your domain name is exmaple.com and your SSH port number is (insert your TCP port number here). You have a development server for developing your own custom ClassicPress theme and you want to use your development server to backup everything from your production server.
Open the terminal and connect to your development server via SSH.
ssh yourusername@devserver -p (your port number if it's other than port 22)
From your development server, create an SSH key pair. We are going to use id_rsa.
ssh-keygen -t id_rsa
Next, enter the location and filename. Example:
/home/yourusername/.ssh/classicpress
After that, leave the passphrase blank. Press Enter a couple of times until you get back to the prompt ending with $.
Execute ssh-copy-id with the name of the public key file and specify the username and domain name.
ssh-copy-id -p (your TCP port number if not 22) -i ~/.ssh/classicpress.pub yourname@example.com
You should be able to login to your server. Give it a try.
ssh -p (your port number or remove -p) -i ~/.ssh/classicpress yourname@example.com
If all goes well, you should be able to connect to your production server without been prompted for the password or passphrase. I mentioned "without passphrase" because if a Linux user executes a single script for performing a backup and has set a passphrase for the SSH identity key, then the script will prompt a Linux user for the passphrase multiple times.
For the purpose of backing up a database, this task will take you through creating a .my.cnf file. This is a hidden file that will contain the username and password for mysqldump command. mysqldump allows a database administrator to backup the MySQL or MariaDB database.
Login to your production server from your development or backup server and perform the steps below.
Create a new file using either vim or nano called .my.cnf. This file will be saved in your home directory. mysqldump will read the file containing the username and password. For me, I use vim.
vim .my.cnf
If you are using Vim, press the i key to begin the INSERT mode and begin typing the following lines. If you are using nano, simply start typing the following.
Replace yourdatabaseusername with your database username and yourdatabasepassword with your database password. When installing either WordPress or ClassicPress on a VPS server, a Linux administrator must have created a database along with the username and set password during the installation process.
Save your changes.
For Vim users, exit out of INSERT mode by pressing the ESC key; then, type :wq to write changes to the .my.cjnf file and quit Vim. The : key begins the command for Vim, w saves the file, and q quits Vim. If you want to not write any changes and quit Vim, then the command is :q!. If the ! were omitted, then Vim will tell you that you need to save your changes before you quit Vim.
For nano users, the keyboard commands for saving the changes and quitting the text editor is CTRL+O for saving changes (press the ENTER key to confirm changes) and CTRL+X to quit the text editor.
Once done, do a database backup of your WordPress or ClassicPress database.
mysqldump ClassicPress > test.sql
This assumes that your database name is ClassicPress. Replace ClassicPress and enter the name of the database that you created when you installed WordPress or ClassicPress in your VPS server.
Your new file called test.sql should be in the same directory that you executed the command for testing. If you open that file up with your chosen editor, you should see all the database commands. Go ahead and close the file.
Log out of your production server by typing exit and press ENTER.
If your database backup is successful, congratulations! That task is done! Your may delete the test.sql file by using the rm command (be careful with that rm command; you might delete files accidentally). The benefit of having a .my.cnf file within the home directory is that you do not want to expose your database password when executing mysqldump. Let's use sleep 10 as an example as the mysqldump command can be very quick once executed.
If you have a Linux machine, open up two terminals and place them side by side.
For the first terminal, execute the following command:
watch 'ps aux | grep sleep'
The watch command will output the ps aux | grep sleep command. Here is the output of the command:
The command sleep 30 will be there for 30 seconds and will disappear from the watch output after the number of seconds have passed.
Use the CTRL+C key to exit out of the watch output. If you are using a Mac, the keyboard command is Control+C. Command+C is for copying text.
My point is, if the mysqldump command gets executed for a long period of time while dumping the entire database, mysqldump can show up in the list of processes. For example, let's say you executed a mysqldump command as follows:
This command will take about 30 seconds when dumping an entire MySQL/MariaDB database. As a result, the output will be as follows (note that this is just an example):
This can be a big problem if an attacker gains access to your server and monitors for the list of processes. That's why it's important to avoid storing passwords in a script whenever possible. That's where .my.cnf configuration file comes in. I did not know about this until I found out about adding a username and password in .my.cnf file. I learn something new almost every single day.
And if you want an example of a real process list, here it is with Apache web server running in my production server:
Okay. That's all for the remote server configuration. Let's get into some real fun part, the configuration of the development server for performing automated backups!
Development or Backup Server
Now here is the script you have all bee waiting for.
#!/bin/sh
# DIRP: Directory path
DIRP=~/cpbackup
# FILE: Partial file name
FILE=$DIRP/classicpress-$(date +%Y%m%d)
# Hostname, IP address, or domain name
HOST=example.com
# Private key for automated logging into an SSH server (no passphrase or password)
PKEY=~/.ssh/classicpress
# TCP Port number (use whatever port you assigned for an SSH server in the
# production server.)
PORT=22
# User name assigned in the remote Linux server
USER=yourusername
# Let's perform some checks. Does the directory in the $DIRP variable exist?
if [ ! -d $DIRP ]
then
echo "Directory not found: $DIRP"
exit 1
fi
# Does the SSH key pair exist?
if [ ! -f $PKEY -a ! -f $PKEY.pub ]
then
echo "SSH key pair $PKEY and $PKEY.pub does not exist. Exiting."
exit 1
fi
# Delete any backup files older than x number of days
find $DIRP -maxdepth 0 -mtime +10 -exec rm {} \;
# Backup the SQL database and store them locally for later restoration.
ssh -p $PORT $USER@$HOST -i $PKEY mysqldump ClassicPress > $FILE.sql
# Next, change directory to /var/www and compress them to standard output
# which then gets redirected to a compressed .tar.gz file.
ssh -p $PORT $USER@$HOST -i $PKEY 'cd /var/www && tar czf - *' > $FILE.tar.gz
# If there is a wp-config.php file stored outside /var/www, make a backup of
# that configuration file as well.
scp -P $PORT -i $PKEY $USER@$HOST:/var/wp-config.php $FILE-wp-config.php
# After that, backup the Apache virtual host configuration file.
scp -P $PORT -i $PKEY \
$USER@$HOST:/etc/apache2/sites-available/000-default.conf $FILE-apache.conf
# The script ran successfully.
exit 0
First, I recommend that you create a bin directory inside your home directory.
mkdir ~/bin
Then, use a text editor in the terminal of your choice (vim, nano, pico, etc.) to create a new file called cpbackup.sh. That script will be in the bin directory. In my case:
vim bin/cpbackup.sh
Copy the script that I created above. It's after the section called Development or Backup Server. The script starts with #!/bin/sh which is the start of the script. Copy it all the way down to exit 0.
Paste the script in the terminal. For Linux users who use a GNOME Terminal like I do, it's CTRL+SHIFT+V. For Mac users who use a Terminal, it's Command+V.
Make some changes to the variables, such as the host name/IP address, port number, et cetera.
Save your changes and exit the text editor.
Give the script an executable permission.
chmod +x bin/cpbackup.sh
chmod, called change mode, allows you to modify read, write, and execute permissions for a user, group, and others. This is beyond the scope of my article. Remember back in the Audience and Prerequisites section that I have to assume you are familiar with Linux. I will have to write another article if I have to get everyone up to speed on how to gain familiar with Linux.
After that, execute the following command:
bin/cpbackup.sh
And you are done! If all goes well, all of your backup files have been stored in the backup directory. And oh, be sure you test your backups by extracting all the WordPress/ClassicPress files from the archive and put it in /var/www. Restoring the database is as simple as:
mysql -u ClassicPress -p ClassicPress < classicpress.sql
Then, simply copy wp-config.php file to /var (it's a good idea to move your wp-config.php file outside of /var/www directory) and copy the Apache configuration file to /etc/apache2/sites-available/, enable the virtual host using the a2ensite command, and you are good to go.
To automatically backup your WordPress/ClassicPress site from time to time, simply execute crontab -e and enter at the bottom of the crontab file:
0 0 * * * bin/whateveryournameofthefileis.sh
And that is done.
Summary
Hopefully you should have a backup infrastructure in place so that if anything goes wrong, you can be able to restore from a good working backup. I hope my article is helpful to anyone who needs to perform a backup of their website including the database. Stay safe and practice good security hygiene online. Oh, and backup your files in your computer to a server or a NAS if you have one. And yes, you should definitely have a home server or a NAS for backing up all your important files. Thank you for reading my article.
Article published: 2022-11-17 08:47
Categories: The World of Computers, Information Technology, Internet, Networking, Scripting and Programming
For those unfamiliar with LXC and Proxmox, LXC is similar to a virtual machine that runs a guest OS (Windows, Linux, Mac, Android, etc.) but the container part of LXC excludes the core part of the OS and simply provides networking and storage inside a container. Unlike virtual machines, applications inside a container can access resources on a host system directly. Proxmox is a hypervisor for running virtual machines and Linux containers (LXC) in a server hardware.
IP Addresses and Subnetting
Do you need to have multiple IP addresses assigned to a network interface inside an LXC container? In terms of networking, a single NIC can have multiple IP addresses. This is useful if you want to run a single server with multiple websites that have their own IP address. Here's what I mean:
172.20.31.0/23
A small HTML file with a list of websites hosted by the web server.
172.20.31.1/23
A web application running Adminer, a lightweight alternative to phpMyAdmin.
172.20.31.2/23
A development version of my website that mirrors a production version. Anyone who visits my site sees my production version of my website.
Once I test the changes I made in the development website, I push the changes up to the production website.
172.20.31.3/23
A custom-built web application for taking notes. Any notes written in HTML gets shown up in the web browser.
172.20.31.4/23
A development version of the note-taking web application taken from 172.20.31.3.
A note for those new to networking: pay attention to the subnet. A slash 23 subnet can start with 172.20.30.1 and ends at 172.20.31.254. Both 172.20.30.255 and 172.20.31.0 are both valid IP addresses. 172.20.30.0 is a network address and 172.20.31.255 is a broadcast address. Cisco has an article about IP addressing and subnetting in order to help you understand how subnetting works. With that out of the way, let's get into configuring a Linux container running in a Proxmox server.
How I Initially Configure the Network Interface Inside a Linux Container?
When you create a new Linux container, you get to the Networking tab and filled in the IP address and default gateway. Once you start the container, the output of /etc/network/interfaces is as follows:
auto lo
iface lo inet loopback
auto eth0
iface eth0 inet static
address 172.20.31.0/23
gateway 172.20.30.1
Now you want to add additional IP addresses to a single network interface:
auto lo
iface lo inet loopback
auto eth0
iface eth0 inet static
address 172.20.31.0/23
gateway 172.20.30.1
iface eth0 inet static
address 172.20.31.1/23
iface eth0 inet static
address 172.20.31.2/23
iface eth0 inet static
address 172.20.31.3/23
iface eth0 inet static
address 172.20.31.4/23
You saved the file and you did a restart of networking service:
systemctl restart networking
What happens if you reboot? The output will be as follows:
If I did not check out the /etc/network/interfaces file, I would execute the ip a command that lists IP addresses for all interfaces and saw something similar to this:
gpadmin-local@webservers:~$ ip a
1: lo: mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0@if111: mtu 1500 qdisc noqueue state UP group default qlen 1000
link/ether 7e:ef:18:b4:a4:b9 brd ff:ff:ff:ff:ff:ff link-netnsid 0
inet 172.20.31.0/23 brd 172.20.31.255 scope global eth0
valid_lft forever preferred_lft forever
inet6 fe80::7cef:18ff:feb4:a4b9/64 scope link
valid_lft forever preferred_lft forever
When I restarted the networking service, I found out that the network interface (eth0) already exists. I don't know why Proxmox would change individual IP addresses to the same IP address and add redundant default gateways with the same IP address, but it seems Proxmox does not handle multiple IP addresses in a network section of the web interface for the Linux container.
Solution
Turns out I have to remove the IP address and gateway from the network configuration. With the Linux container selected (in my case: webservers), I went into Network, double-click the eth0 row containing an IP address, tab over to IPv4/CIDR edit box and removed both the IP address/subnet mask and gateway. Once I've done that, I clicked OK, made configuration changes in /etc/network/interfaces, and rebooted the container. In order to verify full networking functionality, I rebooted the Linux container, logged back into the container, and once I execute ip a command, all the IP addresses are preserved.
A screenshot of the Proxmox web interface. Everything in the sreenshot is darkened except for the two edit boxes, which is IP address/CIDR and default gateway. Both are set to blank and the IPv4/CIDR edit box shows "None." This image only shows a dialog box for network configuration. Those with eyesight can view a large image by clicking or tapping in the figure.
Oh, and I almost forgot: this network configuration can also apply to Ubuntu, AlmaLinux, Rocky Linux, OpenSUSE, and just about any other Linux distributions.
Notes For Those Who Use a Screen Reader
For anyone using a screen reader such as JAWS (Windows), NVDA (Windows), VoiceOver (Mac), or Orca (Linux), I'm not sure how anyone can access a virtual machine within a Proxmox web interface. And I'm not even sure if screen readers can read the text within the Linux container console; however, I am able to copy text from the console. The only difference is I have to highlight using a mouse, right-click in the console, and select copy. Doing Control+C sends a termination signal or displays ^C in a terminal. Doing Control+V once does nothing, but executing Control=V again shows a ^V in a console, so pasting text in a Proxmox console does nothing for those who do not want or cannot use a mouse.
Summary
The bottom line is this: in the networking section of the Proxmox web interface, leave the IP address and default gateway blank. Once you get into the console, edit the /etc.network/interfaces and add your IP addresses manually.
I have been watching a couple of YouTube videos of people who want a computer in one room (such as a wiring closet) and a keyboard, video, and mouse (KVM) in a home office. To give you an idea of what I'm talking about, I want to post links to YouTube videos.
Embedding YouTube or Odysee videos will insert a tracking cookie in users' personal computers. As a citizen of the US, I need to follow GDPR if European visitors visit my website. I don't like and want to talk to lawyers to be honest. 🤣😀
As for the video from Linus Tech Tips, I would much rather have a couple of computers rather than single computer that can house a couple of virtual machines running desktop OSes such as Linux and Windows just to make it easier for me. So yeah, a virtual machine is a computer within a computer that can serve different purposes such as running Ubuntu within Windows using VirtualBox or by running Windows OS in a Linux host using KVM or Xen.
So, I have an idea of my own. When it comes to building a house, I would like to wire my future house for Ethernet and HDMI connectivity. The computer will be in a wiring and server closet and my essentials such as my mouse, keyboard, monitor, audio interface, and a couple of others will be in a recording studio/home office room. So here's what I'm thinking of buying in the future:
One concern I had with USB over Ethernet is latency from the audio interface to the computer over the Ethernet cable. However, one look at the images and answers to questions tells me that latency over Ethernet cable should not be a factor. That way, I can have my computer fans spin at max RPM in the server room (well, maybe not too loud) and still have complete silence in my studio/office with sound proofing and acoustic panels.
Oh, here something that I would like to show you (for those with eyesight). It's a home theater room made in Blender.
This is a rendering of my home theater made in Blender. It has a 200" projection screen, 7.2.4-channel speaker system, and comfy seats with cupholders in between.
My dream of a home theater will be a lot simplistic than that with flat ceilings with no light strips in between them. A home theater will be in a basement. Speaking of home theater, I could patch my home desktop over to my home theater while in the server closet if I want to watch YouTube/Odysee videos. How cool would that be if a house can have a central computer core just like in a starship such as U.S.S. Enterprise D or U.S.S. Voyager? Speaking of starships in Star Trek, do you know that a central computer core can span over several decks? Actually, a starship can contain two or more computer cores. My house might have only a single central computer core with a couple of rack-mounted desktops and servers as part of a homelab! Here is an example of a homelab shown in a YouTube video.
As for rack-mounted desktop PCs, I'm thinking of Rosewill RSV-L4500U Rackmount Server Chassis and put them inside a StarTech.com 42U Server Rack Cabinet. Oh, sure the rack-mounted enclosure costs more than an open-frame 42U rack, but to me, it will look a lot cleaner by having an enclosure. Now if only there are 5U or 6U rack-mountable computer cases so I can fit a tall heatsink such as Noctua NH-D15 or Scythe Ninja 5 heatsink. Now keep in mind. I do not care about the looks of computer components. No RGB fans, no tempered glass side panels from the likes of Lian-Li 011 Dynamic XL or even from Phanteks line of computer cases, no nothing.
Why!?
Out of sight, out of mind.
That's my mindset when it comes to computers. If you watch Star Trek, you must know that Captain James T. Kirk does not look at a central computer core every single day while he's in his quarter or in the bridge! No Starfleet officers should care about how cool computers look as long as they perform their jobs! Sorry elite gamers, but I do not like the market the computer case manufactures are targeting. Go look at how cool your computer looks while you are failing your Cisco CCNA course because hey! All that cool RGB fans and that NZXT Kraken Z73 cooler of yours are so much fun to look at! 🤣
As for me? I'm excelling in Cisco CCNA course! Why? My Silverstone FARA R1 computer case only has a plain side panel and is free of distractions. It's unfortunate that a computer case with a plain side panel is unavailable. But I have it! However, the limitation I have is I cannot take the case feet off so I can fit on top of a rack-mounted shelf.
Oh, did I meant to taunt you for having all that cool tempered glass computer case with all that RGB gizmos? Get over it! Seriously. I know you've been playing games 24/7 while you look at how cool your computer looks. Okay, okay. I'll be nice to you elite gamers! 🙂
If you want to get into the world of homelab, Lawrence Systems and Learn Linux TV has done a couple of videos for setting up different aspects of homelab such as Home Assistant, storage server, firewalls and switches, Linux, Ansible, securing your lab, and so much more. Check out the playlist on YouTube and enjoy! Any desktop PCs will be part of a homelab as well. And yes, RAID is not a backup solution. In a RAID 5 setup with 7 hard drives (6 for data and 1 for parity), if one drive fails, swap out the hard drive as soon as possible. But if two drives out of 7 fail, you lose all the data inside the hard drives.
Now here's a question. What about entertainment devices such as NVIDIA Shield TV, and PlayStation 5? These devices can sit on a rack-mountable shelf and can be patched to my home office room using an Ethernet-based USB extender. For HDMI, I can make use of Anthem MRX 740 for my home office room and use HDMI Zone 2 Out to pass HDMI audio/video signal to Anthem MRX 1140 receiver, which can go to the home theater room. What I've learned from a thread over at avforums.com is that the main receiver (MRX 740) will pass audio/video signal to another receiver and plays no part for processing A/V signals.
What that means is I can independently select a source in the second zone while the first source will remain intact in the main office room. Just tell Home Assistant to switch to NVIDIA Shield in the second zone of MRX 740 receiver, and Home Assistant will do the rest. Whatever is shown in MRX 1140 for the home theater room will reflect what was shown from MRX 740's HDMI Zone 2 output.
So yeah, centralizing desktop PCs in a wiring/server closet is my dream of the future! Centralize all the desktop computers into one 45U server rack! 🙂
Article published: 2021-08-14 17:42
Categories: Visionary Living and Exploring Tomorrow, Homes and Buildings, Home Theater, The World of Computers, Computers, Networking
When you setup your new router, it's always a good idea to create a new user other than admin for pfSense and vyos for VyOS in order to reduce the chance that bots and miscreants will gain access to your router.
VyOS
Here's the completed configuration of my VyOS router and I will show you the commands.
Configuration
service {
# ...
ssh {
access-control {
allow {
user <username>
user vyos
}
}
listen-address 10.249.0.1
}
}
system {
# ...
login {
banner {
pre-login "Unauthorized access is strictly prohibited."
}
user <username> {
authentication {
encrypted-password ****************
plaintext-password ****************
}
full-name "First and last name goes here."
home-directory /home/<username>
}
user vyos {
authentication {
encrypted-password ****************
plaintext-password ****************
}
}
}
# ...
}
Commands
ssh vyos@10.249.0.1
configure
edit system login user <username>
set authentication plaintext-password <your-password-goes-here>
set full-name "First and last name goes here."
set home-directory /home/<username>
exit
edit service ssh access-control
set allow user <username>
set allow user vyos
commit
save
You want to allow vyos access using SSH to make sure it works. Also, there is encrypted-password in VyOS but VyOS gave me an error telling me that the encrypted password is invalid. I did try to discard, but VyOS told me there are not changes to be discarded, so I saved, started a new terminal window, and once I SSH into my VyOS router for 10.249.1.1, everything works fine.
Now don't exit out of VyOS session just yet. You want to make sure SSH is working properly for a user you want to log into. Because otherwise editing and viewing the configuration will have to be done either through the use of a console cable or a monitor and keyboard hooked up to a monitor. SSH using your new username and password you've created. If you can successfully login to VyOS with a different username, you can simply remove the vyos user from the access control list in configuration mode.
delete service ssh access-control allow user vyos
Again, stay logged in to VyOS and use a different terminal to test and make sure you can log into VyOS through SSH. If everything is working as intended, you can safely log out of VyOS from all the terminals you've opened.
Also, you can configure a banner. Examine the configuration above and see if you can add a login banner. The pre-login is for when a user attempts to access the VyOS router using SSH. This will print out a banner before a user gets prompted for a password. After a user logs into VyOS, if the post-login is set, VyOS will print out the banner once the user logs in. This concludes the commands used for securing VyOS.
pfSense
The same can be done for pfSense. Open the web browser, point your browser to pfSense (in my case, http://10.249.2.1), and login to your pfSense web interface. Once you get to the main interface, follow instructions as follows.
In the System menu, open the User Manager.
Click in the + Add button below the list of users.
Enter the Username, Password, and Full Name. No spaces in the username.
In the Group Membership area, select admins and click in Move to "member of" list. This will move the admins group to the "member of" list.
Save the changes, log out, and log back in as the new admin user you have created in step 4.
In the user manager, click in the pencil icon (Edit) to edit the admin user.
Check the checkbox for Disabled. An admin user cannot login once the checkbox is selected.
When done, Save the changes.
Try to login as admin. If successful, you should not be able to log in as an admin user but instead log in as a new user. This concludes the step-by-step instructions for pfSense.
Conclusion
Preventing a root or admin user from logging into a router is one of the security's best practices. You can help ensure that bots and miscreants won't be able to gain access to your router without the correct username and password. Even when bots are performing a brute-force attack. Still, it's important to restrict access to the router through the use of a management subnet and if using pfSense, setup a root and server certificate in the Cert. Manager within the System menu and add a root certificate to your web browser of your choice. Use a management subnet for any devices that have SSH access or a web interface and do not allow managers, sales, web developers, or any other non-IT departments access to the critical network infrastructure.
Update: I just hit "c" twice in my keyboard (ccode instead of code) even though I only typed "c" just once. Ugh... Maybe I just need a different keyboard that prevents double-types regardless of the operating system I'm using... (And yes, I'm using Arch Linux.)
In part 1 of the article, I have covered how to create and configure virtual machines and install both VyOS and pfSense. In part 2 of this article, I'm going to cover how to configure pfSense with OSPF networking. Let's get going, shall we?
You must have created and configured virtual machines, installed VyOS and pfSense, and setup VyOS with OSPF. This article is for the intended audience who have knowledge and good experience with installing operating systems, working with Linux command line, basic networking, and virtual machines. This article is for those who have good computer networking skills such as configure networks and how to segregate networks from each other. Also, it's important to know how subnetting in IPv4 works.
I'm going to spin up a new virtual machine in my desktop computer. The virtual machine will have a network bridge set to ethbr0_2402. This is what the image looks like as shown below.
There are four terminals that shows how I configure all the interfaces.
For those who have eyesight, in the upper-right hand corner of the screen, you can see a list of bridges I have created. Each bridge has its own VLAN ID. So in my case, ethbr0_2402 has a VLAN ID of 2402. The traffic between the desktop and my server is as follows:
From desktop...
Signal coming out of virtual machine in my desktop...
...goes to ethbr0_2402, a bridge interface, which goes to...
...to ethbr0.2402, a VLAN interface that assigns a VLAN ID tag of 2402, which goes to...
...ethbr0, another bridge interface, which carries tagged VLAN traffic to...
...enp5s0, a physical network interface that carries tagged VLAN traffic to...
...an RJ-45 Ethernet cable which carries bits of ones and zeros of VLAN traffic to...
...my server, which gets picked up by enp7s0, which sends tagged VLAN traffic to...
...ethbr10, a bridge interface which sends tagged VLAN traffic to...
...vlan2402, a VLAN interface which links to ethbr10. vlan2402 removes the VLAN tag from the signal and then goes to...
...ethbr10_2402, a bridged interface which now sends traffic to pfSense VM in my server.
I hope I don't lose everyone. That's why it's important to have a lot of experience in Linux and networking. In order to send traffic back to a desktop VM running in my desktop, you simply reverse the process. Here, let me put it this way:
Desktop } Desktop VM <-> ethbr0_2402 (Bridge) <-> ethbr0.2402 (VLAN) <-> ethbr0 (bridge) <-> enp5s0 (Physical NIC in Desktop) <-> Ethernet Cable <-> enp7s0 (Physical NIC in Server) <-> ethbr10 (Bridge) <-> vlan2402 (VLAN Interface) <-> ethbr10_2402 (Bridge) <-> pfSense VM { Server
If VLANs did not exist, I would have to have 10+ network interface cards or adapters for my desktop and server (that makes 20 total or 2 times the number of network interfaces cards or adapters) and 10+ network cables and that would be a lot of mess to deal with.
And yes, do refer to the scenario and netplan configuration for my server in my first article.
Let's Configure pfSense
A Note About The Quality of Images
Please excuse the quality of images. When I took screenshots using a virtual machine (From the menu: Virtual Machine -> Take screenshot), the images were saved as PNG format (Portable Network Graphics). The PNG format is larger in file size than JPG (jpeg) format and I wanted to do a batch conversion of PNG to JPG all at once. In Linux, that's where mogrify comes in and is included in a package called imagemagick that can be downloaded from the distro repository such as Ubuntu, Debian, Fedora, Arch, OpenSUSE, and many others. I'm not completely sure how to preserve the image quality at 80% while keeping the file size low. Plus, I also used mogrify to batch resize images and store it in a different folder so that files would not be overwritten. I do have GIMP but because I have about 39 images, it can take a lot of time to do repetitive tasks.
For the first tab, I do Ctrl+Shift+E and change from .png to .jpg.
Resize the image (Alt+I, S).
Do Ctrl+Shift+E again to export the new image inside the thumbnail folder.
Because I cannot do Ctrl+Tab to get to the next tab, I had to click in the second tab while magnified, because in a 4K screen (3840x2160), the tabs are very small. That's because GIMP is not HIDPI-aware in Linux.
I then would repeat the same process 38 times until I have gotten all the images converted and resized, making sure that the image quality is at 80%, not 90% by default in GIMP.
I don't know why I cannot do Ctrl+Tab and Ctrl+Shift+Tab for switching between images without using my mouse. This is why I would rather want to do automation using a command line utility whenever possible and that's the reason why I couldn't figure out how to set the quality of images using the mogrify command. So, with a note about image quality out of the way, let's get into configuring pfSense for networking, shall we?
Images for Configuring pfSense
Figure 1-0: Login to pfSense. Username is "admin" and password is "pfsense." All lower case.Figure 1-1: IntroductionFigure 1-2: Configure hostname, local domain name, and DNS server.Figure 1-3: Set your time zone. Leave the NTP server as default unless you want to use your own.Figure 1-4: Leave the WAN configuration unless you want to change the IPv4 address. Scroll down to the bottom of the page...Figure 1-5: ...and clear the ckeckbox for bogon and RFC 1918 networks. This is only if you are behind another router.Figure 1-6: Leave the LAN configuratin as it is.Figure 1-7: Choose your password.Figure 1-8: Reload configurationFigure 1-9: Done!
Instructions for Configuring pfSense
Open the web browser for the virtual machine, type the pfSense's internal IP address, and hit Enter.
The virtual machine should have the same VLAN ID as the pfSense's virtual machine for the internal network. in my case, the IP address for the pfSense gateway is 10.249.2.1 and the VLAN ID is 2402. Note that you do not choose a VLAN interface when configuring virtual machines, but bridges. The network bridge is what allows virtual machines to share one or more interfaces.
The default username is admin and the password is pfsense. Login to pfSense. (Figure 1-0)
The setup wizard begins. Continue to the next step. (Figure 1-1)
Enter the hostname for pfSense, your domain name (optional; I've chosen lab.graysonpeddie.lan because this is my "virtual homelab"), primary/secondary DNS server (I suggest 1.1.1.1 and 1.0.0.1, but you can pick one such as 8.8.8.8 or 9.9.9.9), and I suggest you uncheck the checkbox for Override DNS. (Figure 1-2)
Leave the NTP (Network Time Protocol) Server as it is, and choose a timezone if you want. (Figure 1-3)
For the next step, leave everything as is, scroll down to the bottom of the page and uncheck the two check boxes for blocking private and bogon networks. These two need to be unchecked for Internet connection to work correctly. Continue to the next step. (Figure 1-4, Figure 1-5)
Note: If you are configuring pfSense and you have a public IP address assigned by your Internet service provider, leave both of them checked.
This step is for configuring a private IP address. Leave it as it is and continue to the next step. (Figure 1-6)
Choose your own password. Make sure you can remember your password as you'll need it when you log into pfSense. (Figure 1-7)
Click Reload to reload the configuration and click Finish to finish the setup wizard.
Things to Do Before Configuring OSPF in pfSense
Images for Post-Installation Steps and Verifying Internet Connection
Figure 1-1: Hover over System menu and go to Advanced.Figure 1-2: Go to Networking section.Figure 1-3: Disable hardware offloading and click Save.Figure 1-4: If all goes well, you should have connectivity to the Internet.
Let's Take Care of Things Before Vefifying Internet Connectivity
Skip the steps below if you are configuring pfSense in a real hardware. These steps are for those working in virtual machines.
Open the System menu and click in Advanced. (Figure 1-1)
Go to the Networking section and scroll down to the bottom. (Figure 1-2)
At the bottom of Networking section, check the checkbox to disable hardware checksum offloading. (Figure 1-3)
Verify that the Internet is working in the internal network. (Figure 1-4)
Install FRR Package in pfSense
Images for Installing FRR Package in pfSense
Figure 1-1: Available Packages.Figure 1-2: Click Install next to FRR.Figure 1-3: Click Confirm to install FRR.Figure 1-4: The installation of FRR should be successful.
How to Installing FRR Package in pfSense
Go to the Package Manager that is inside the System menu.
In the Available Packages section, scroll down to frr and click Install. (Figure 1-1, Figure 1-2)
Confirm the installation of FRR package and the installation should be successful. (Figure 1-3, Figure 1-4)
pfSense Firewall
pfSense must allow OSPF traffic from two VyOS routers. It's important to limit OSPF traffic between pfSense and VyOS routers. Let's configure the firewall before we configure OSPF using FRR.
Images for Firewall Configuration
Figure 1-1: Firewall Rules for WANFigure 1-2: Allow OSPF traffic.Figure 1-3: New firewall rule for OSPF added in WAN.
How to Allow OSPF Traffic Through the WAN Firewall
Open Rules in the Firewall menu. You should be in Firewall / Rules / WAN. The WAN tab is selected by default. (Figure 1-1)
Click in the Add button (either one) to create a new rule. (Figure 1-1)
For a new rule, the action is set to Pass by default. Set the Protocol to Any, set the source IP address and subnet mask (remember to limit the subnet for OSPF traffic to pass through), and click Save at the bottom of the page. In my case, I changed the source drop down menu to Network and set the address to 172.24.19.0/28. (Figure 1-2)
The firewall rule configuration is done. (Figure 1-3)
Now Let's Configure OSPF
Images for OSPF Configuration
Figure 1-0: The FRR-related menu items can be found in the Services menu.Figure 1-1: Enable FRR in order for OSPF to work.Figure 1-2: Enable OSPF and set the Router ID.Figure 1-3: Set the default area to 0. Note that I have added IP addresses for networks, but because that section has been marked depricated, I've decided to remove the IPv4 addresses from the list of networks.Figure 1-4: It's time to add a new interface.Figure 1-5: Add a WAN interface and set the interface to "broadcast."Figure 1-6: Add a LAN interface and set the interface to "non-broadcast."Figure 1-7: OSPF configuration is complete.
Instructions for OSPF Configuration
Under the Services menu, click in FRR Global/Zebra. This will take you to the global configuration page for FRR. (Figure 1-0)
Check the checkbox for Enable FRR and click Save at the bottom of the page. (Figure 1-1)
Next, go to the OSPF section and enable OSPF. Specify the Router ID. (Figure 1-2)
Scroll down until you get to specify an area. The area should be the same as configured in VyOS routers. Save changes when done. (Figure 1-3)
Note that there is no need to specify the networks which is marked deprecated. Interfaces need to be specified instead.
Go to the Interfaces tab and click Add. (Figure 1-4)
Set the Interface to WAN, Network Type to Broadcast, and save changes. (Figure 1-5)
Add another interface for LAN, with Network Type set to Non-Broadcast and save. (Figure 1-6)
You should have both the WAN and LAN interfaces configured for OSPF. (Figure 1-7)
Can Hosts Behind pfSense Ping Other Hosts in Different Networks?
Now, if you go ahead and start up two hosts from two VyOS networks (one host per network), you should be able to ping a host behind pfSense (In my case, 10.249.2.100). But before we do that, let's check the neighbors.
Let's start with vyos_2400. I'm going to connect to the VyOS router via SSH, issue sh ip ospf neigh, and get the list of neighbors from there.
grayson@v2400-host1:~$ ssh vyos@10.249.0.1
Welcome to VyOS
vyos@10.249.0.1's password:
Last login: Sun Feb 21 21:42:53 2021 from 10.249.0.100
vyos@v2400-vyos:~$ sh ip ospf neigh
Neighbor ID Pri State Dead Time Address Interface RXmtL RqstL DBsmL
10.249.1.1 1 Full/DR 34.369s 172.24.19.3 eth0:172.24.19.2 0 0 0
10.249.2.1 1 Full/DROther 34.372s 172.24.19.4 eth0:172.24.19.2 0 0 0
vyos@v2400-vyos:~$ exit
logout
Connection to 10.249.0.1 closed.
grayson@v2400-host1:~$
So, my vyos_2400 router can see pfSense's "public" IP address. Can I ping a host behind pfSense's network?
grayson@v2400-host1:~$ ping 10.249.2.100
PING 10.249.2.100 (10.249.2.100) 56(84) bytes of data.
64 bytes from 10.249.2.100: icmp_seq=1 ttl=62 time=1.22 ms
64 bytes from 10.249.2.100: icmp_seq=2 ttl=62 time=2.09 ms
64 bytes from 10.249.2.100: icmp_seq=3 ttl=62 time=1.20 ms
64 bytes from 10.249.2.100: icmp_seq=4 ttl=62 time=1.90 ms
^C
...
Yes I can! The ^C is for Ctrl+C which terminates the process. What about a host behind vyos_2401? What do the neighbors look like?
grayson@pop-os:~$ ssh vyos@10.249.1.1
Welcome to VyOS
vyos@10.249.1.1's password:
Last login: Wed Feb 10 21:57:29 2021 from 10.249.1.100
vyos@v2401-vyos:~$ sh ip ospf n
Neighbor ID Pri State Dead Time Address Interface RXmtL RqstL DBsmL
10.249.0.1 1 Full/Backup 37.358s 172.24.19.2 eth0:172.24.19.3 0 0 0
10.249.2.1 1 Full/DROther 32.289s 172.24.19.4 eth0:172.24.19.3 0 0 0
vyos@v2401-vyos:~$ exit
logout
Connection to 10.249.1.1 closed.
grayson@pop-os:~$
Well, I don't see a need for ping as it should work with no problems. But it should work nonetheless.
Okay, so what about a host behind pfSense? Can I ping one of the hosts behind VyOS routers from a host that is behind pfSense? I tried and the answer to that is "no." The reason for that is NAT needs to be configured so that pfSense knows how to return the traffic back to the original host. pfSense needs to translate the original source address (such as 10.249.0.100) back to 10.249.2.0/24, which is the translation address and subnet mask. This is where outbound NAT comes in. Note that hosts behind pfSense can reach the Internet and pfSense will translate the packets back to the WAN interface address, but for private networks via OSPF, this is where the outbound rule needs to be filled in.
Network Address Translation (NAT)
NAT can be found in the Firewall menu. Instead of giving you images, I can provide step-by-step along with the list of settings that need to be configured.
Here are the list of settings that need to be configured.
Advanced Outbound NAT Entry
Interface: WAN
Address Family: IPv4
Protocol: Any
Source: Network; 10.249.2.0/24
Destination: Network; 10.249.0.0/16
Translation
Address: Other Subnet (Enter Below)
Other Subnet:10.249.2.0/24
Note that your IP addresses may be different depending on how you configure your routers and networks. Adjust accordingly.
Now here's step-by-step instruction on how to configure NAT.
Once you get to the NAT page that is found under the Firewall, go to the Outbound section.
In the Mode area, click in Hybrid Outbound and click Save.
Below the Mappings section, click in the Add button. You will be adding a single rule for OSPF so either buttons do not matter. The order of rules do matter for both firewall and NAT rules.
Enter the settings I have provided above. If you are using a screen reader, jump to the previous heading level 4 titled Network Address Translation, and jump down to the first list.
Once you are done configuring the rule from the settings I've provided, save your changes and click Apply at the top of the screen. pfSense will apply the changes.
As I understand it, the "source" field in the NAT rule is for hosts originating behind pfSense and the destination field is for hosts that is behind the two VyOS routers. And the translation address is where pfSense will send the return packets back to the originating hosts behind pfSense. Maybe I could configure the same thing in one of my VyOS routers? Why would I do that while I could simply use masquerade as a translation address? Why is there no such thing as masquerade in pfSense? That's the question I'm going to find out for myself.
So Now Can Hosts Behind pfSense Ping Other Hosts in Different Networks?
With all the settings I have applied for my NAT rule, the host behind pfSense can now ping hosts that are behind VyOS routers. To verify, I have deleted the rule with a translation address of 10.249.2.0/24, and tried pinging hosts behind the VyOS routers. The echo replies did not get sent back to the originating host behind pfSense. I re-added the rule again with the settings from above and pings from other 10.249.0.0/16 networks are now working.
Conclusion
That is all for part 2 out of 2 for this article! The hosts between pfSense and VyOS routers can ping each other once OSPF has been configured! If you enjoyed my article, can you please share your feedback with me in Twitter and LinkedIn? I will provide links so you can comment in my article. Thank you for reading my article. I've been writing an article since February 10th and stopped since the 14th as I've been busy with my studies during the last two weeks and until the 21st of this month. Happy networking! Have fun!
Revisions
Version 1.00:Initial Publish
Version 1.01:Updated the figures for the pfSense's setup wizard and OSPF configuration.
If you understand computer networking and know how IPv4 subnetting works, here's a zip file which contains a self-contained HTML file. Double-click in the HTML file and you can begin practicing.
This is part 1 of 2 of configuring multiple networks that can communicate with each other through OSPF.
Introduction
How much do you know computer networking? Do you know how subnetting works? What about IP addresses? Do you know how routers and switches work? Do you have a homelab and do you know what a homelab is? If you answer yes to all of the questions and you want to expand your knowledge of networking, this article is for you. Yes, I'm targeting audience that have a good knowledge in networking. This is even for those with lack of certificates such as CompTIA A+, Network+, and Security+, and even for those without a degree! Well, why don't we delve right into it, shall we? If you are Network+ certified, you must know that OSPF is a dynamic link-state protocol that allows the two or more private networks to talk to each other. If you have a consumer router such as Netgear or Linksys, this article is only for the pros!
Also, my article covers the use of virtual machines and networking bridging, so I'm going to assume you know how to set them up. I'm using Ubuntu Server 20.10 as my Linux home server that runs KVM (Kernel-based Virtual Machine). Virtual machines are what enables a computer to run inside a computer and network bridging behaves similar to a network switch. And because of that, I'm also going to assume you are familiar with the Linux command line.
Now buckle your seatbelt because this article is going to be a very long one.
Internet Protocol version 4; a 32-bit addressing scheme in a form of a.b.c.d (0 to 255 for each octet)
OSPF
Open Shortest Path First is a dynamic link-state protocol.
NAT
Network Address Translation: used to provide public-to-private IP address translation when an ISP can only assign a single public IP address. IPv6 (128-bit addressing scheme does away with NAT because each device can be assigned a public IPv6 address).
KVM
Kernel-based Virtual Machine for running virtual machines in a home server or desktop. A virtual machine is akin to putting a computer inside a computer each running its own operating system.
Router
A router connects one or more networks together and even connects the network to the Internet. A router can act as a default gateway if an IP address falls outside the list of network the router is connected to.
Switch
A network forms a star topology by connecting more than one host (a computer, printer, home theater receiver, etc.) to a central switch. If an IP address falls outside the subnet, a switch can forward a packet to a router.
You can learn a lot about computer networking by going through the tutorial. Basic networking is beyond the scope of my article.
You work at a company called "GalaxyTech" that specializes in building PCs and servers. Your company has three buildings. The two buildings connect to a central building through fiber. The central building has a fiber connection to the Internet Service Provider. Each of the three buildings have a purpose:
Central building: Office rooms, IT department, web development, sales department, and accounting.
Manufacturing facility: for building computers and servers, has rooms for office spaces, and IT department.
Warehouse building: warehouse, shipping, office rooms, and IT department.
Your job is to travel to three buildings and get networking setup. An ISP has provided a set of IP addresses within a subnet, which is 172.24.19.0/29. For the purposes of the article, I'm using 172.24.19.0/29 as a private subnet. 172.16.0.0/12 is part of RFC 1918 private addressing scheme. Plus, the three networks are going to be as follows:
Central building: 10.249.0.0/24
Manufacturing facility: 10.249.1.0/24
Warehouse building: 10.249.2.0/24
The three networks should be able to talk to each other. Note, however, this is going to be a very simple scenario. In the production environment, it is important to split a single subnet into a smaller subnet and segment them into different networks. Plus, different networks should not talk to each other and pinging should not be allowed for security reasons. But in this article, I am keeping it simple.
My computer and server both have 2.5Gbit Ethernet adapters installed. There is no switch involved as I cannot afford a managed 8-port switch that supports VLANs at the time of writing, so I simply have a straight cable that connects from my computer to my server. And because I'm running Ubuntu Server 20.10 in my home server, I want to share how my home network is configured (/etc/netplan/00-installer-config.yaml).
# This is the network config written by 'subiquity'
network:
version: 2
ethernets:
enp3s0f1:
dhcp4: no
enp4s0f0:
dhcp4: no
enp4s0f1:
dhcp4: no
enp7s0:
dhcp4: no
bridges:
...
ethbr10: # Computers, Smartphones, etc.
interfaces:
- enp7s0
- enp4s0f0
- vlan10
dhcp4: no
addresses:
- 172.20.1.1/26
ethbr10_2400:
interfaces:
- vlan2400
dhcp4: no
ethbr10_2401:
interfaces:
- vlan2401
dhcp4: no
ethbr10_2402:
interfaces:
- vlan2402
dhcp4: no
...
ethbr10_2409:
interfaces:
- vlan2409
dhcp4: no
ethbr10_2419:
interfaces:
- vlan2419
dhcp4: no
ethbr10_2429:
interfaces:
- vlan2429
dhcp4: no
ethbr20: # Home Automation, Wired
interfaces:
- enp4s0f1
dhcp4: no
addresses:
- 172.20.5.1/24
...
vlans:
vlan10:
id: 10
link: enp3s0f1
dhcp4: no
...
vlan2400:
id: 2400
link: ethbr10
dhcp4: no
vlan2401:
id: 2401
link: ethbr10
dhcp4: no
vlan2402:
id: 2402
link: ethbr10
dhcp4: no
...
vlan2408:
id: 2408
link: ethbr10
dhcp4: no
vlan2409:
id: 2409
link: ethbr10
dhcp4: no
vlan2419:
id: 2419
link: ethbr10
dhcp4: no
vlan2429:
id: 2429
link: ethbr10
dhcp4: no
My home server is equipped with a 4-port Gigabit networking card, a built-in NIC in my motherboard, and a 2.5Gbit networking adapter. The motherboard's built-in NIC is not in use. My computer's IP address is 172.20.1.8, which is part of a /26 subnet. I do have a couple of VLANs setup in my desktop computer as well. But instead of Ubuntu, I'm running Arch Linux, which is for experienced Linux users such as myself. So back to the Netplan config file, Netplan seems to be exclusive to ubuntu and is not available in other Linux distributions such as CentOS and Debian. All of the interfaces have DHCP set to no so this is more like a router that connects multiple networks together and pfSense is a router distribution that runs in a virtual machine. pfSense has an IP address of 172.20.0.2 and resides in ethbr0 bridge interface which is not shown here in order to keep it short. My computer, laptop, and smartphone resides in ethbr10. Each VLAN has an ID and is linked to a specific bridge interface, which in this case, ethbr10. Then, each bridge is specifically bridged to each VLAN. So, the way it works is this: ethbr10_2400 is linked to vlan2400 with a VLAN ID of 2400, vlan2400 is linked to ethbr10, and ethbr10 is bridged to multiple interfaces including a VLAN interface with an ID of 10. Sure, my networking is very complex, but that is part of the homelab!
As for my desktop computer, I use Network Manager for networking. Here's what it looks like:
There are four terminals that shows how I configure all the interfaces.
In order to create virtual machines, I wanted to do so as a user instead of root. To do that, I added a user called kvmguests and added libvirt as a group.
The acl package provides the setfacl and getfacl. For those who do not know, ACL stands for "Access Control List" which grants or denies read/write/execute permissions for files and directories. /vm/kvm contains two directories: iso and img. The iso directory is for holding installation images and img directory is for holding virtual machine disk images.
This image shows the list of storage pools for the virtual machines shown in the right. To the left shows how to open the "Details" window.
So, to get to the list of storage pools, right-click where it says "QEMU/KVM," click Details, and in the Connection Details dialog box, click Storage. You can add whatever storage pool you need. I suggest keeping the installation images separate from disk images. This helps keeps it organized.
For VyOS, you'll want the rolling release as the LTS release requires a purchase of a subscription.
Ready to Configure Virtual Machines for OSPF?
Okay, so you got your virtual machine environment and your network all setup and ready to go, so it's time to get your virtual machines up and running. For those who have eyesight, you can click in the checkbox to show the list of images. For blind users, I do not know how this is going to work for you when it comes to exiting out of the virtual machine window using your keyboard. You could make use of serial connection by deleting the video display when you customize the installation before starting a virtual machine. However, I've ran into problems when trying to get the pfSense installer image up and running. Please accept my apologies for not being of much help if you rely on your screen reader. Anyway, let's get going.
Images for Installing and Configuring Virtual Machines
Figure 1-0: The prerequisite for installing a virtual machine. Connect to a remote server via SSH.Figure 1-1: The first step of installing a virtual machine. Choose local install by default.Figure 1-2: The second and third step of installing a virtual machine. Choose an install image.Figure 1-3: The fourth and fifth step of installing a virtual machine. Choose Debian 10 for VyOS or FreeBSD 11.4 for pfSense 2.4. 1 Gig of RAM and 2 virtual CPUs are fine for testing.Figure 1-4: The sixth and seventh step of installing a virtual machine. Create a disk image.Figure 1-5: The eighth step of installing a virtual machine. Make sure to name, select "Customize configuration before install," and select a network bridge.Figure 1-6: The ninth step of installing a virtual machine. Add a network interface.Figure 1-6: The tenth step of installing a virtual machine. Let the installation begin!
Prerequisite
First, enable virtualization extensions in your computer's BIOS. This will depend on what motherboard/PC/laptop you use. If you can't see the screen, ask someone to help you navigate around the menus in the BIOS screen.
If you do not have virt-manager installed in your desktop, use your package manager to install virt-manager.
Debian and Ubuntu:apt update && apt install virt-manager
Arch Linux:pacman -Sy virt-manager ("S" stands for "sync" and "y" stands for "update repositories.")
If you have not done so, create a private/public key pair with a blank passphrase so that you can login to a remote server without a password. A Virtual Machine Manager will prompt you for a password indefinitely. Also, make sure your user is part of the libvirt group. Because otherwise connecting to a remote server using virt-manager is not possible.
Open the terminal and type the following command.
ssh-keygen -t rsa
Accept the defaults and leave the passphrase blank.
Then enter the following command with your username and hostname/IP address:
You will be prompted to enter the password in order to copy a public key to your remote server. Never share your sensitive private and public key with everyone.
Try to login to your remote server.
ssh kvmguests@vmserver
If you can login without a password, you are good to go.
Open virt-manager and connect to your remote server (figure 1-0). With a public key, you should not be prompted for a password.
Go to File, then Add Connection... to open the Add Connection window.
Check the checkbox for Connect to remote host over SSH.
Enter the Username and Hostname.
Then click Connect.
Okay, so you are able to login to the remote server using virt-manager, right? Let's get into the fun part, which is creating a new virtual machine!
With the remote server highlighted, click in the Create a new virtual machine button. A tooltip will come up if you hover over a monitor with a play button in the toolbar. (Figure 1-1)
The Local install media is selected. Click Forward to go to the next screen. (Figure 1-1)
To the right of Choose ISO or CDROM install media, click the Browse button to the right of the edit box. (Figure 1-2)
In the list of storage pools, choose a pool that contains all hte ISOs you've downloaded. In my case, it's ISOs. (Figure 1-2)
In the list of ISO files, look for vyos-rolling-latest.iso. Then click Choose Volume to confirm your selection. (Figure 1-2)
In the Choose the operating system you are installing edit box, type Debian 1 and use the up and down arrow keys to select Debian 10. VyOS is based on Debian. Then click Forward. (Figure 1-3)
Accept the Memory and CPUs settings and click Forward. (Figure 1-3)
Select or create custom storage and click Browse. (Figure 1-4)
Select a storage pool where the disk images will be in (in my case, Images) and click the plus button next to Volumes to Create New Volume. Hover the mouse pointer over the plus sign for the tooltip to show up. (Figure 1-4)
Give the storage unit a name. Here is the convention and example name that I use. (Figure 1-4)
(VLAN ID for external network)_(VLAN ID for internal network)_(Operating System Name)
v2419_v2400_vyos
v stands for VLAN. VLAN 2419 will be the bridge that connects to pfSense (my edge router, a gateway to the Internet). VLAN ID 2400 is for an internal bridged network that VyOS VM will use.
Leave the rest as is and click Finish. Make sure you have plenty of disk space in your server! (Figure 1-4)
Once you are back in the New VM, click the Forward button to continue to the last step for the wizard. (Figure 1-4)
Give your virtual machine the same Name as you created in Step 10. Keep the convention the same. (Figure 1-5)
Check the checkbox for Customize the configuration before install. (Figure 1-5) A virtual machine will have two NICs (Network Interface Controllers).
Expand the Network selection, choose Bridge Device... from the dropdown menu, and in the Device name: edit box, assign the bridge to the first NIC. In my case, it's ethbr10_2419. (Figure 1-5)
Remember that VLAN 2419 is for the external network that connects to pfSense (172.24.19.1). The VM will have an IP address of 172.24.19.2/24. That will be for the central building of GalaxyTech. Read the scenario above if you need to.
Be sure you have Customize the configuration before install checked and click Finish. We're not done yet. (Figure 1-5)
Click in the Add Hardware at the bottom left-hand corner of the VM window. (Figure 1-6)
In the list of hardware components, select Network. This is where a second NIC can be added. (Figure 1-6)
Next to the Network source:, open the dropdown menu, choose Network bridge, and enter the name of the bridge in Device name. I chose a bridge for an internal network, which in my case it's ethbr10_2400. 2400 is my VLAN ID. (Figure 1-6)
When done, click Finish to return to the VM window. (Figure 1-6)
Notice that you have two NICs in the list of hardware components installed for the VM. (Figure 1-7)
Now, it's time to Begin Installation. Click in that button in the upper-left corner and on we go!!!
Figure 1-0: Booting up the installation of VyOS: Press Enter to accept the default.Figure 1-1: Installing VyOS: The username and password is vyos. Begin the installation of VyOS.Figure 1-2: Installing VyOS: Type ? to see the list of commands. Accept the defaults for partitioning the disk.Figure 1-3: Installing VyOS: Accept defaults and create a password for the user vyos.Figure 1-4: Installing VyOS: For VM, type vda to install the bootloader into the disk. When done, reboot.Figure 1-5: Booted up VyOS for the first time! Login and you're in.
Install VyOS Into The First Instance of VM
When the bootloader shows up, leave the Live (amd64-vyos) as selected and press Enter. (Figure 1-0)
When you get a login prompt, the username and password is vyos. (Figure 1-1)
Type the following command to begin the installation of VyOS. (Figure 1-1 and 1-2)
install image
Then press Enter. Note that you can type ? to see the list of commands and parameters. (Figure 1-2)
The installer has found a disk to partition in callsed vda. Go ahead and accept the defaults (auto and vda). (Figure 1-2)
Input Yes to continue the installation as the disk is empty. (Figure 1-3)
You can leave everything as default until you get to create a password. Note that I should have left the image name as-is. An image is an image similar to Cisco IOS; however, I did enter v2419_v2400_vyos. Just leave the image name as default. (Figure 1-3)
Enter the password that you can easily remember. (Figure 1-3)
Once you are done creating a password, leave the rest as defaults and reboot. (Figure 1-4)
Once rebooted, login with the user vyos along with your newly-created password and you're in!
Congratulations! You've just created yourself a first virtual machine and installed VyOS! Before you continue configuring the network, let's create another virtual machine and install VyOS.
I have already provided instructions for creating a virtual machine and installing VyOS into the VM. In Step 10 and Step 13, give your new disk and VM the new name while still following the convention.
(VLAN ID for external network)_(VLAN ID for internal network)_(Operating System Name)
v2419_v2401_vyos
Your VM will use a different internal bridge name for the second VLAN, which has an ID of 2401. So, in Step 19 (third from the last step), you would enter ethbr10_2401 for the second NIC. So, here's how I have it setup:
VM Name
External NIC
Internal NIC
v2419_v2400_vyos
ethbr10_2419
ethbr10_2400
v2419_v2401_vyos
ethbr10_2419
ethbr10_2401
v2419_v2402_pfSense
ethbr10_2419
ethbr10_2402
And the IP addresses and subnet prefixes are going to be:
VM Name
Public IP
Private IP
v2419_v2400_vyos
172.24.19.2/24
10.249.0.1/24
v2419_v2401_vyos
172.24.19.3/24
10.249.1.1/24
v2419_v2402_pfSense
172.24.19.4/24
10.249.2.1/24
Note that I simply wanted to pretend that anything in 172.24.19.x/24 is part of the public IP addressing scheme. 172.24.19.x/24 is still part of 172.16.x.x/12 private subnet (RFC 1918). Only for this article, I'm going to pretend that anything in 172.24.19.x/24 will have a public IP address assigned to each of the three virtual machines. A fake one, if you will.
Images for the installation of pfSense into the third VM
Figure 1-1: Installing pfSense: Accept copyright notice.Figure 1-2: Installing pfSense: Accept choices made by the installer.Figure 1-3: Configure pfSense: Configure connection to the Internet.Figure 1-4: Configure pfSense: Setup the internal network and ping 1.1.1.1 to confirm the Internet is working.
Create a Third VM and Install pfSense In a VM
I am very sure you can create a new VM named v2419_v2402_pfSense with an internal network bridge named ethbr10_2402. If not, look over the steps on how to create a new VM. Be sure to keep ethbr10_2419 as the device name for the first NIC in the New VM dialog.
Now the steps for installing and configuring pfSense is as follows:
Go ahead and press the Enter key throughout the installation process. The installer will insteall pfSense into the virtual disk and prompt you to reboot the system. (Figure 1-1 and 1-2)
Set vtnet0 as WAN and vtnet1 as LAN. (Figure 1-3)
pfSense will attempt to discover an IP address if a DHCP server is configured. If not, wait for pfSense to time out. (Figure 1-3)
When you get to the menu, type to select 2 and press Enter. (Figure 1-3)
Enter 1 for vtnet0. Type n as DHCP server was not configured. (Figure 1-3)
Enter the IP address for external network. In my case, it's 172.24.19.4/24. If I omit /24 from the IP address, I will be asked to input a subnet bit (or a prefix number). 24 is what I would enter. (Figure 1-3)
If you have no need for IPv6, simply type n, hit Enter, and hit Enter again to leave the next line blank.
Now here's something I screwed up on. Not shown in Figure 1-3 is it asks if you want to revert back to HTTP instead of keeping in HTTPS. Idealy, for a production environment you always want HTTPS for protection against on-path attacks, which is also known as man-in-the-middle (MiTM) attack. Because this is a homelab and I have kept the networks segregated from each other, HTTP is fine.
Press the Enter key and you should be back on the menu. (Figure 1-4)
Press 2 again and press Enter to get into the interface configuration. You'll want to configure the LAN address which is number 2.
If you only want IPv4, simply enter the IP address and subnet mask (in my case, 10.249.2.1/24) and leave the rest set to default. Press Enter at the last step to get back to the menu.
To make sure everything is working, press 7, hit Enter, and ping (don't type ping) 1.1.1.1. If it's working, you are good to go.
Congratulations on installing and setting up pfSense. But it's not over yet. You will need to setup another virtual machine to configure pfSense. But let's get the network up and running in two VyOS instances first.
If you are familiar with Cisco's IOS, then the commands will be a lot different. First, and foremost, there's no such thing as privileged EXEC mode. The $ sign means you are in user mode. If you type sudo -i or configure, VyOS will put you in "#" mode. Please note the following output below:
vyos@vyos:~$ sudo -i
root@vyos:~# exit
logout
vyos@vyos:~$ configure
[edit]
vyos@vyos# hostname v2400-vyos
hostname: you must be root to change the hostname
[edit]
vyos@vyos# set hostn?
Configuration path: [hostn] is not valid.
[edit]
vyos@vyos#
VyOS is a full Linux distribution. Every VyOS command begins with "set" once you are in VyOS configuration mode. Even though VyOS is a Linux distribution that is focused in routing, please do not edit the /etc/hostname file as root. Instead, perform the following commands:
vyos@vyos:~$ configure
[edit]
vyos@vyos# set system host-name v2400-vyos
[edit]
vyos@vyos# commit
[edit]
vyos@vyos# save
Saving configuration to '/config/config.boot'...
[edit]
vyos@vyos# exit
exit
vyos@vyos:~$ reboot
Are you sure you want to reboot this system? [y/N] y
Reboot the system and you should be back in VyOS with a new hostname. Set the hostname for the second VyOS instance (different VM) to v2401-vyos and reboot the VM. The commit command commits changes to memory and save saves changes to the disk as /config/config.boot. VyOS will complain that changes made to /etc/hostname and then executing the two set commands above will cause VyOS to error out saying "temporary failure in name resolution" or something like that. So yea, unlike Cisco IOS's hostname v2400-cisco in global configuration mode, changes in VyOS do not take effect even once the changes have been commited to memory and saved to disk. However, just about all the networking commands you'll be doing will take effect once you call commit in configuration mode.
See what happens if you configure a domain name for your VyOS router. You can name whatever domain name you want. I suggest you keep the TLD set to lan so you won't cause confusion with a .com domain.
configure
set system domain-name lab.graysonpeddie.lan
commit
discard
save
set system domain-name lab2.graysonpeddie.lan
discard
The first command gets you into configuration mode. I then configure the domain name for the router. I commited the changes to memory but decided to discard the changes. Because the change has been stored in RAM, I get an error message saying No changes have been discarded. Then, I changed the domain name to lab2.graysonpeddie.lan but decided to discard the changes. So even if I make changes to network configuration, the changes won't take effect until I commit to RAM and then save changes to disk.
Now, let's get into networking, shall we? First, we must configure the interfaces for v2419-v2400-vyos. Make sure you are in configuration mode. There should be no :~ between hostname and number sign. Remember, sudo -i takes you into root mode just like all other Linux distributions do.
set interfaces ethernet eth0 address 172.24.19.2/24
set interfaces ethernet eth1 address 10.249.0.1/24
set protocols static route 0.0.0.0/0 next-hop 172.24.19.1
commit
save
And for v2419-v2401-vyos VM:
set interfaces ethernet eth0 address 172.24.19.3/24
set interfaces ethernet eth1 address 10.249.1.1/24
set protocols static route 0.0.0.0/0 next-hop 172.24.19.1
commit
save
Once you exit the configuration mode for both of them, can you ping 1.1.1.1 in both VMs? If yes, you are in good shape! 0.0.0.0/0 means a default gateway of last resort. 1.1.1.1 is not in the routing table, so the router has to go through 172.24.19.1 in order to reach out to 192.168.1.254, which is my mom's router, and 192.168.2.254, which is the Comtrend router for Consolidated Communications. Because I do not have a way of getting traceroute information from a virtual machine, I can do that from my desktop PC. Note that I cannot ping 172.20.0.2, which is the edge router for pfSense that reaches out to the Internet. I don't want my internal hosts to ping my edge gateway for security and obscurity reasons.
[grayson@grayson-epcotcenter Storage]$ traceroute 1.1.1.1
traceroute to 1.1.1.1 (1.1.1.1), 30 hops max, 60 byte packets
1 _gateway (172.20.1.1) 0.191 ms 0.143 ms 0.136 ms
2 * * *
3 192.168.2.254 (192.168.2.254) 4.231 ms 4.246 ms 4.267 ms
4 sdsl-blt-216-227-16-10.gtcom.net (216.227.16.10) 14.952 ms 14.960 ms 14.976 ms
5 rblt-aa.gtcom.net (216.227.16.1) 15.049 ms 15.080 ms 15.023 ms
6 atl-b22-link.telia.net (62.115.9.130) 24.028 ms atl-b22-link.telia.net (80.239.192.68) 23.200 ms 23.194 ms
7 cloudflare-ic-309901-atl-bb1.ip.twelve99-cust.net (213.248.83.18) 53.332 ms cloudflare-ic-301665-atl-bb1.c.telia.net (62.115.32.222) 43.354 ms 43.360 ms
8 one.one.one.one (1.1.1.1) 32.600 ms 24.107 ms 21.845 ms
[grayson@grayson-epcotcenter Storage]$ ping 172.20.0.2
PING 172.20.0.2 (172.20.0.2) 56(84) bytes of data.
^C
--- 172.20.0.2 ping statistics ---
4 packets transmitted, 0 received, 100% packet loss, time 3055ms
[grayson@grayson-epcotcenter Storage]$
For learning experience, let's configure DHCP server for both VyOS routers. I'm going to start with v2419-v2400-vyos VM in configuration mode.
edit service dhcp-server shared-network-name InternalNetwork2400 subnet 10.249.0.0/24
set range P1 start 10.249.0.100
set range P1 stop 10.249.0.199
set default-router 10.249.0.1
set name-server 1.1.1.1
commit
save
Here's another way without the use of edit:
set service dhcp-server shared-network-name InternalNetwork2400 subnet 10.249.0.0/24 range P1 start 10.249.0.100
set service dhcp-server shared-network-name InternalNetwork2400 subnet 10.249.0.0/24 range P1 stop 10.249.0.199
set service dhcp-server shared-network-name InternalNetwork2400 subnet 10.249.0.0/24 default-router 10.249.0.1
set service dhcp-server shared-network-name InternalNetwork2400 subnet 10.249.0.0/24 name-server 1.1.1.1
commit
save
I'll let you figure out why I use the edit command as a shortcut.
And for v2419-v2401-vyos VM in configuration mode:
edit service dhcp-server shared-network-name InternalNetwork2401 subnet 10.249.1.0/24
set range P1 start 10.249.1.100
set range P1 stop 10.249.1.199
set default-router 10.249.1.1
set name-server 1.1.1.1
commit
save
P1 is just a text field for the range. Think of it as a pool of IP addresses. This allows VMs to acquire an IP address from the router that connects to it.
NAT
Endpoint devices using private IP addresses need to communicate over the Internet over a public IP address. Configure NAT in both VyOS virtual machines. Remember, this can only be done in configuration mode.
set nat source rule 1 translation address masquerade
set nat source rule 1 outbound-interface eth0
commit
save
Setting a translation address to masquerade allows endpoint devices to act as though it has its own public IP address.
Now this is where OSPF comes into play. But first, I will create two virtual machines--one of which connect to their own network behind VyOS routers. In order to do that, though, I will have to set up my virtual machine environment in my desktop machine.
[grayson@grayson-epcotcenter ~]$ sudo usermod -a -G libvirt grayson
[sudo] password for grayson:
[grayson@grayson-epcotcenter ~]$ sudo systemctl enable libvirtd
Created symlink /etc/systemd/system/multi-user.target.wants/libvirtd.service → /usr/lib/systemd/system/libvirtd.service.
Created symlink /etc/systemd/system/sockets.target.wants/virtlockd.socket → /usr/lib/systemd/system/virtlockd.socket.
Created symlink /etc/systemd/system/sockets.target.wants/virtlogd.socket → /usr/lib/systemd/system/virtlogd.socket.
Created symlink /etc/systemd/system/sockets.target.wants/libvirtd.socket → /usr/lib/systemd/system/libvirtd.socket.
Created symlink /etc/systemd/system/sockets.target.wants/libvirtd-ro.socket → /usr/lib/systemd/system/libvirtd-ro.socket.
For changes to take into effect, I have to log out and back in for the libvirt group to take effect for my user grayson. But instead of logging out and back in, I'm going to restart my machine. I could even execute a start command for libvirtd if I want. libvirtd daemon (similar to a "service" in Windows) is what allows me to run virtual machines in my desktop without root privileges.
Once I restart my desktop machine, I am able to spin up 3 virtual machines, install Pop!_OS, and ping the router's IP addresses. My desktop had 32GB of RAM and my server has 48GB of RAM. For my desktop, I can give the three virtual machines each having 4GB of RAM. Of course, I cannot ping each other's internal endpoints because each router (be it VyOS or pfSense) does not know how to reach their own network that came from a different router.
While we can set static routes for each VyOS router as we did for setting a default gateway, we are going to make use of OSPF. More information about OSPF can be found in this web page. Actually, the web page I linked can show you how to configure Cisco routers using OSPF, but I am only focusing in VyOS and pfSense for now.
Let's learn how we can configure OSPF in two of the VyOS instances. Let's start with v2419-v2400-vyos and then configure v2419-v2401-vyos.
vyos@v2400-vyos# edit protocols ospf
vyos@v2400-vyos# area 0 network 10.249.0.0/24
vyos@v2400-vyos# area 0 network 172.24.19.0/28
vyos@v2400-vyos# parameters router-id 10.249.0.1
vyos@v2400-vyos# commit
vyos@v2400-vyos# save
vyos@v2401-vyos# edit protocols ospf
vyos@v2401-vyos# area 0 network 10.249.1.0/24
vyos@v2401-vyos# area 0 network 172.24.19.0/28
vyos@v2401-vyos# parameters router-id 10.249.1.1
vyos@v2401-vyos# commit
vyos@v2401-vyos# save
Each router has two networks in the same area and the router ID is the neighbor ID. Let's exit the configuration mode and have a look at the neighbors.
vyos@v2400-vyos# exit
exit
vyos@v2400-vyos:~$ sh ip ospf neighbor
Neighbor ID Pri State Dead Time Address Interface RXmtL RqstL DBsmL
10.249.1.1 1 Full/Backup 37.035s 172.24.19.3 eth0:172.24.19.2 0 0 0
vyos@v2401-vyos# exit
exit
vyos@v2401-vyos:~$ sh ip ospf neigh
Neighbor ID Pri State Dead Time Address Interface RXmtL RqstL DBsmL
10.249.0.1 1 Full/DR 33.747s 172.24.19.2 eth0:172.24.19.3 0 0 0
Pay no attention to the last three columns including the dead time (actually, you can read more about dead time that I linked here). The two VyOS routers can see each other. Each router has an outbound interface which is eth0 and has its IP address assigned. From 10.249.0.0/24, to get to 10.249.1.0/24, the router goes out from 172.24.19.2 to 172.24.19.3 in order to reach the 10.249.1.0/24 subnet. So, from my virtual machine running Pop!_OS that connects to v2400-vyos, my VM's IP address is 10.249.0.100 and when I ping 10.249.1.100 the ping is successful.
I then ping from another virtual machine that is connected to v2401-vyos VM. Its IP address is 10.249.1.100. I ping 10.249.0.100 and if you look at the neighbor for v2401-vyos, the ping is successful.
So everything is successful between two VyOS instances, right? Let's configure the firewall next.Update as of October 24, 2021
Oh, and here is something I forgot to mention. It's about setting passive interfaces. None of the hosts inside the private network should know anything about OSPF traffic. That's where the next command comes in.
set protocols ospf passive-interface eth1
That way, VyOS will not broadcast OSPF into the internal network. What if you have 5 networks? Let's say there are 4 LAN networks and 1 network that connects to other OSPF-enabled routers such as pfSense and Cisco ISR routers. You can use the set of commands instead as shown below.
set protocols ospf passive-interface default
set protocols ospf passive-interface-excluse eth0
All the interfaces are set to passive by default except the one that connects to the OSPF network and out to the Internet.
Why is firewall important? It's to keep the bad guys out and control which traffic can go in and out. Let's configure the firewall for both VyOS VMs.
First, I will show you the configuration.
vyos@v2400-vyos# show firewall
name WANIn {
default-action drop
rule 1 {
action accept
state {
established enable
related enable
}
}
rule 2 {
action accept
source {
address 172.24.19.0/28
}
}
rule 3 {
action accept
source {
address 10.249.0.0/16
}
}
}
name WANLocal {
default-action accept
rule 1 {
action accept
state {
established enable
related enable
}
}
rule 2 {
action accept
protocol ospf
}
}
[edit]
vyos@v2400-vyos#
vyos@v2401-vyos# sh firewall
name WANIn {
default-action drop
rule 1 {
action accept
state {
established enable
related enable
}
}
rule 2 {
action accept
source {
address 172.24.19.0/28
}
}
rule 3 {
action accept
source {
address 10.249.0.0/16
}
}
}
name WANLocal {
default-action drop
rule 1 {
action accept
state {
established enable
related enable
}
}
rule 2 {
action accept
protocol ospf
}
- rule 3 {
- action accept
- destination {
- port 22
- }
- protocol tcp
- }
}
[edit]
vyos@v2401-vyos#
I will start with the commands for v2400-vyos and you can go from there.
edit firewall name WANIn
default action drop
...
But before you enter the commands above and figure out the rest, let me ask you this. What does a dash indicate?
If you guess "the changes will be deleted before it's commited to RAM," you are correct. This is how to delete the firewall rule:
delete firewall name WANLocal rule 2
Now go ahead and perform the rest of the commands. Follow the hierarchy!
Did you complete all the commands? Okay! More questions for you! How do you assign the two sets of firewall chains to the outbound interface?
Do you know how you configured IP addresses for your interfaces? Use the ? and TAB key to help you out.
set interfaces ethernet eth0 firewall in name WANIn
set interfaces ethernet eth0 firewall local WANLocal
What does WANIn and WANLocal do?
The firewall controls the flow of traffic coming into the router (WANLocal) or into the network (WANIn). However, the firewall chain for out has not been configured. This is fine for what we need to setup OSPF.
Can you ping each other's public IP addresses (172.24.19.x)? If no, what do you need to do in order to allow the public IP addresses to ping each other? I'll give you a hint...
We have installed VyOS into virtual machines, set our passwords, configured Ethernet interfaces, NAT, OSPF, and firewall. In my two desktop virtual machines, I can ping each other's IP addresses in different networks.
I have to end it here because the article got so long and it took me a few days to write this article. In Part 2, I'm going to finish off by configuring pfSense that will allow internal network to communicate with other networks behind VyOS routers. Stay tuned!
I Welcome Your Feedback!
Instead of enabling comments in my blog, I encource you to give feedback in one of the following links below.
Version 1.01:Added info to the prerequisite section for those who are blind.
Version 1.02:I need to add an internal network (10.249.0.0/16) as a source in the WANIn chain for both VyOS routers. Without it, even though I am able to ping each other's endpoints from different networks, I cannot ping the two VyOS internal networks from a host behind pfSense. However, once I add the third rule to a WANIn chain, I'm able to ping hosts behind the two VyOS routers. I also added the feedback heading and revisions to the table of content.
Version 1.03:I have added a PNG image that shows a multiple routers that are connected to the central router and the central router connects to the "cloud."
Version 2.00:(October 24, 2021) The developers of VyOS must have changed from dns-server to name-server when it cones to configuring a DHCP pool. The changes should work now. Plus, I've added information about passive interfaces, which is important for keeping the OSPF traffic from broadcasting into the private network.
